Less time spent on documentation thanks to AI, more time for patient care: what does the scientific evidence say?
In February 2025, a study published in JAMA Network Open added compelling evidence to the growing role of artificial intelligence in addressing clinician burnout and documentation overload. Researchers at the University of Pennsylvania evaluated an ambient AI scribe tool in a 7-week pilot across 46 outpatient clinicians representing 17 specialties within a large academic health system. By comparing electronic health record (EHR) metrics before and during implementation, the team was able to quantify the tool’s real-world impact on clinical workflow.
Measurable efficiency gains
The results demonstrated consistent improvements:
- Around 15–20% less time spent per note
- A 17–30% reduction in after-hours documentation (“pajama time”)
- Higher same-day visit closure rates
- Reduced typing volume, reflecting automation of documentation tasks
These objective EHR metrics were reinforced by clinician feedback. Many participants reported a lower cognitive burden, describing less mental fragmentation during consultations and improved ability to maintain eye contact and active listening.
Beyond productivity: restoring attention to patients
Importantly, the benefits extended beyond time savings. Clinicians described improved engagement during encounters, with less need to divide attention between the patient and the computer screen. In a healthcare environment where documentation demands continue to expand, this shift is not trivial, it directly affects care quality, physician satisfaction, and potentially patient trust.
That said, the study also highlighted current limitations. AI-generated notes still required review and occasional editing for accuracy, completeness, and stylistic alignment with individual practice patterns. Reliability and seamless integration remain critical areas for refinement.
From innovation to infrastructure
Taken together, the findings suggest that ambient AI scribes are moving from experimental tools to foundational clinical infrastructure. By reducing administrative friction and reclaiming cognitive bandwidth, these systems may play a meaningful role in mitigating burnout while preserving clinical focus.
As documentation pressures grow worldwide, the question may no longer be whether AI belongs in clinical documentation, but how to deploy it responsibly, securely, and at scale to maximize both efficiency and human presence in care.
Source: JAMA Network Open